Force-Feeding Taxonomy to the Enterprise

By Christian Buckley posted 08-28-2012 01:32


While presenting at last year's Taxonomy Bootcamp (a KMWorld event) in Washington DC, I was able to attend a number of sessions on information architecture and taxonomy development given by people outside of the SharePoint space. It's always interesting to hear different perspectives from across various platforms, and compare notes -- especially when the learning comes from people with PhDs in the areas of library sciences. Most people in the SharePoint space tend to only attend SharePoint events, and that's unfortunate. Not surprisingly, most of the issues I heard are common across all collaborative platforms: problems with end user adoption, inability to define and enforce adequate governance policies and procedures, and problems getting management support (and funding) for planning activities, such as defining and implementing taxonomy up front, rather than retroactively.


Many organizations do not understand the importance of metadata and keyword taxonomies to the health of the platform, and their role in the overall success of SharePoint. I run into this again and again as I talk to customers and attendees to various events. Management teams push back on recommendations to spend time and resources on up-front planning, not understanding the far greater cost of implementing later, when in production. This is one area that falls within that old project management truism: if you fail to plan, you are planning to fail. We do not take these fundamental steps in planning out our SharePoint environments, and then we wonder why, months down the road, people are not using the platforms, and end users complain that they can't find their content.


What came out of the Taxonomy Bootcamp loud and clear was that metadata is tied to everything that happens within any knowledge management platform, including SharePoint.


<soapbox> Metadata is what defines our content, puts it into context for collaborating across teams, it is what drives our social interactions, and it is what powers workflow and other forms of automation. It is the fundamental building block of every collaborative platform. Metadata can be structured (taxonomy) and unstructured (folksonomy, or end user-generated keywords). Social tools utilize it, and add to it, which in turns improves the overall search experience. You cannot search (and expect to find anything) without it. And yet simply adding to it is not the answer -- you must actively manage it, massage it, cut and trim it. Metadata management is not static, but an ongoing activity. </soapbox>


Most organizations have two problems with explaining the importance of metadata management to their leadership teams: how to quantifying the ROI of the planning activity, and once approved (usually at a later stage where things are more dire), figuring out an action plan and critical roles to complete the plan. Both of these are complex activities, and there is no quick and easy action list that will adequately cover everything involved and all possible scenarios. But since we are a culture of Top 10 Lists and one hundred and forty character responses, I'll take a stab at it:

  • The Return on Investment (ROI) of planning your taxonomy strategy
    The string of benefits from having a strong taxonomy is tied to the search experience: a strong taxonomy allows end users to more accurately tag their content, which improves the search experience, which leads to improved productivity, which results in increased usage of the platform, which leads to faster realization of the financial investment you've made in the platform. While most organizations do not track performance data on search results, there is hard data available online showing that improved taxonomies and classification have greatly enhanced research and productivity efforts. The real discussion should happen around the opportunity cost: would you rather spend 2 to 4 weeks of planning now (and the associated resource costs) or have to do a large portion of that planning later PLUS the time it takes to implement changes to your information architecture, roll out new content types, and map expanded term sets to hundreds, thousands, or tens of thousands of documents? There's only so much you can automate -- this is a very manual process, which is why it makes sense to do as much of it up front as possible. If you take some time to estimate the time it will take per document to apply the new taxonomy, figure out the cost of that person's time, and then multiply that number by the total number of documents in your system, the number is likely staggering….and may be compelling enough to get your CFO to change his or her mind about funding your planning budget now.
  • Best practices for building and implementing your taxonomy strategy
    There is no single method for achieving taxonomy nirvana. So much depends upon your corporate culture and the methods you use to work with your end users and across business units to drive your IT projects. Generally, there is a governance body of some kind, responsible for compliance and risk management, of which SharePoint is just a piece of the larger pie. This group will guide your efforts around the policies and procedures, the company and government standards and regulations (depending on your industry), and the competing IT project initiatives around which you must work. Building a taxonomy is an iterative activity, requiring a team effort to define and expand upon the model, run through accepted use cases, and share the results of your planning with all stakeholders for feedback. Once implemented, your strategy should include methods for ongoing feedback (such as enabling end user tagging -- or folksonomy -- as a way of learning through use, as well as by providing a solid change management mechanism) and proactive management of the taxonomy. Someone should be given the task of reviewing and updating the taxonomy on a weekly basis (at least at first) so that end user suggestions can be quickly added.


I'm just scratching the surface with both of these topics, and would love to hear some of your best practices. Have you come up with a winning model for explaining the ROI of proactive planning, and successfully illustrating the cost of doing nothing? Do you have best practices that you've employed within your own organization for building and implementing a taxonomy strategy. Please share what has worked for you. 

#metadata #InformationGovernance #Taxonomy #sharepoint #governance #buckleyplanet #Collaboration #SharePoint #TaxonomyandMetadata


09-12-2012 10:27

This post really reflected many of the same issues and challenges I've faced as an Information Management practitioner trying to integrate better practices and value-based outcomes for business clients and colleagues in SharePoint 2010. I have said many of the same things myself.
I have found some success with two things in particular: using the term "tags", and focusing on separating the good from the bad. Since "metadata" is often seen as jargon, I found that making reference to hash tags or grocery store aisles is often understood better by non-subject matter experts -- it's also becoming part of our everyday chatter, so it's not so foreign. Likewise, instead of talking about retention schedules and disposition periods, I have focused the discussion on being able to quickly and easily separate what you want to keep from what you want to get rid of. Thankfully, we've got some real allies on the tech side who support making the connection betwteen retaining everything and poor system performance (e.g., slower response times); when we keep more than we need to, we slow everything down -- that's a message that resonates more than one about policies.
Specifically, I am encouraging the use of the Managed Metadata Service (e.g., Term Store) in conjunction with well-defined Content Types to control content. Creating a look-up column in the Content Types to call the MMS gives users a quick and easy way to tag and keep going (and it offers a type-ahead feature). We're still in early implementation, but it's the best-balance solution I've been able to find so far. Plus, Content Types work well with the Key Filters and search facets work out of the box to help refine searches and find relevant content fast.
Remember -- there are two ways to find a needle in a haystack: with a magnet, or with a match. Give users a choice, make it easy for them to tag, and you'll be on your way to the magnet approach.

09-07-2012 00:48

Christian, I completely agree that organizations that fail to plan, plan to fail. One of the great challenges is getting people to understand complex terms like taxonomy and metadata. These are big words and while they are accurate descriptors of library science concepts, they are often just mumbo jumbo to business users; the ones who actually need this stuff to do their jobs.
As you've pointed out, taxonomy and metadata are critical to provide relevant search results, deliver informative reports and trigger workflow. In the delivery of our content management applications, we've found it useful to focus on questions such as:
- How do you want to search for content? By what search terms? By what filters or sorts?
- What do you want the report to look like? What are the column and row labels?
- What are the status changes or values that should kick off or trigger a process?
We allow the answers to determine what the taxonomy or metadata should be but avoid discussing the terms directly as eyes glass over as users struggle to truly understand the concepts. Search more effectively; derive more intelligence from reports; automatically kicking off downstream activities are things business users get excited about. ROI will take care of itself at that point.

08-28-2012 09:29

Excellent research and post